Skip to content
Video

Viggle Remake Turbo

4 шага вместо 6 для замены людей и правки сцен

2 min readupdated

Комьюнити-репак от akhaliq для Viggle Remake: 4-шаговый рендеринг через стек LoRA Remake и SMD MiniMax-H3 Ref2VA 4-step. Модель работает с video-to-video: заменяет людей и редактирует сцены в существующих видео. База — MiniMax-H3, формат — LoRA для ComfyUI.

Ключевые особенности

  • Четыре denoising-прохода вместо шести — примерно в 1,5 раза меньше оценок модели на рендер; по карточке, на треть быстрее поверх int8/GGUF-ускорений Viggle.
  • Два варианта: merged LoRA одним файлом (Remake + SMD слиты во fp32 и пересобраны через SVD до rank 256) или последовательная загрузка двух LoRA — математически то же самое.
  • Есть отдельные merged-файлы для режимов swap и edit, а также готовые API-workflows для ComfyUI.
  • Важно: LoRA нужно грузить через LoraLoaderBypassModelOnly, а не LoraLoaderModelOnly; также используйте diffusion-файл fl2va, не ref2va.
  • 4-шаговая сетка sigma задаётся одним ManualSigmas: 1.0, 0.997228, 0.990826, 0.96, 0.0; euler, BasicGuider без CFG, RemakeKeyframeUntil на 0.9.

Пример

cd ComfyUI
hf download akhaliq/viggle-remake-turbo --include "*.safetensors" "workflows/*" --local-dir ./remake-turbo
cp remake-turbo/*.safetensors models/loras/
# затем запустить workflows/remake_turbo_swap_api.json или remake_turbo_edit_api.json
ПодробнееСкорость: 6 → 4 denoising forwards, ~1,5× меньше оценок модели за рендер.
Compact-профиль базы в примере — около 41 GB (int8/nvfp4); bf16-файлы работают так же.
Merged LoRA: слияние во fp32 и SVD до rank 256.
Лицензия — MiniMax H3 Community License: условия redistribution и commercial use, включая письменное разрешение при выручке выше $20 млн/год.
Запускайте только на людях, которые согласились, и помечайте результат как AI-generated.
Discuss Telegram 5

Will it run?

Working out the memory needs

The bot is still collecting this model's weight sizes; the estimate will appear here in a few minutes. Meanwhile you can rent a GPU with headroom.

Massed ComputeOpen Massed Compute

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # akhaliq/viggle-remake-turbo
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download akhaliq/viggle-remake-turbo --local-dir ~/models/viggle-remake-turbo
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to viggle-remake-turbo
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • SageAttention 3v2.2.0from October 28, 2025

    even faster on RTX 50 / Blackwell

    FP4 attention for Blackwell. Faster than SageAttention 2, but the authors suggest v2 when accuracy matters.

    How to enable: Build from the sageattention3_blackwell folder of the repo; in ComfyUI via KJNodes.

    Earlier generations (2)
    • SageAttention 2++ — Quantised (INT8/FP8) attention with no visible quality loss. The main speed-up for image and video generation.
    • SageAttention 1 — First version: INT8 for Q·K. Still works, but 2++ is faster and more accurate.
  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (3)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
    • xFormers — The older way to speed up attention, works almost everywhere. Sage or Flash are faster where available.
  • cache-dit (DBCache)v1.5.6from November 20, 2026new release

    1.5–3× at nearly the same quality

    Reuses block outputs between diffusion steps. Supports most DiT models.

    How to enable: pip install cache-dit; one line cache_dit.enable_cache(pipe) for diffusers.

    Earlier generations (2)
    • MagCache — Skips steps by magnitude of change. More accurate than TeaCache, no calibration.
    • TeaCache — The first popular diffusion step cache. Works, newer methods are more accurate.
  • LightX2V step distill0.6.1from November 9, 2026new release

    5–10× fewer steps

    Distilled LoRAs and an engine: video in 4–8 steps instead of 30–50.

    How to enable: Load their step-distill LoRA, CFG 1.0, 4–8 steps.

  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

MiniMax H3 prompts

All prompts

Ready prompts with the resulting video: copy and try.

Model family

New posts about versions of viggle-remake-turbo (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

akhaliq/viggle-remake-turbo
Author
akhaliq
Category
Video
Released
Posted

On NeyroLearn

Hugging Face

05

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments